The sanity check for copying socket default value was using the socket
idname for type compatibility checks, which is too strict. Subtypes
would not be recognized as copyable, but since only the plain data type
is needed below this is all that needs to be checked.
That alone would not cause crash (just missing default value copy),
but the tree view template was messing with the default_value DNA
directly by freeing it in advance, which is not necessary and should
be left to blenkernel when freeing the node. Otherwise this would
leave the node invalid without a default_value if the copy function
bails out.
become invalid because the render database is not immediately freed after the job finishes. To avoid access to dangling node pointers, store the function callback in the exec data itself. The node
pointer must not be accessed in the free function (wasn't used before either), these functions are purely for the execution data.
The internal connections used for muted nodes are a "best guess" only. The algorithm would connect linked input/output sockets first, regardless of sensible connections. Replaced this by a more strict
connection method now which only uses matching types, so that Normal outputs won't get input from values, etc.
This is still far from perfect though. Eventually it may be necessary to implement dedicated methods for certain types of nodes to respect different semantics - using the plain socket type is just not
enough information. Also the value used for connections from muted nodes with no matching input will currently be the hidden(!) socket input values, this can lead to unexpected results. Needs further
thought ...
actually just point to existing defined stack values. To make the link check work, set the socket->link pointer if internal node connections exist, this works for both muted and reroute nodes.
necessary.
Problem was/is that the nodes can set the 'type' property, but in order to actually change the data type they would also need to update the typeinfo pointer (and idname), which is strongly discouraged.
Solution is to just replace the input/output sockets of the reroute node with new sockets of the desired type and port all links over.
texture coordinate that should automatically use the default normal or texture
coordinate appropriate for that node, rather than some fixed value specified by
the user.
Issue was caused by ntreeUpdateTree calling for a ntree
which is not in G.main.
This lead to issues in ntreeVerifyNodes (which is called
from ntreeUpdateTree).
Made is so ntreeUpdateTree now accepts main as an argument.
Will work for the release, later we could either solve the
TODO mentioned in ntreeUpdateTree which will eliminate need
in main there or make it so context's main is used from all
over where ntreeUpdateTree is called (currently there're
still some usages of G.main).
When nodes are loaded from a .blend file they can potentially have undefined types. This can happen if a type has been deprecated and removed, or if node types were defined in a python script that has not been loaded correctly. Previously all such nodes would automatically be removed from a node tree, assuming that their types were deprecated and no longer in use (more commonly caused by loading new nodes in an older Blender version). Due to the possibility of dynamic registration it is no longer feasible to simply delete such nodes.
Display and handling of node trees was simply disabled before this patch, so that a node tree where any node or socket type was undefined would not be displayed at all. To give more information and avoid problems caused by necessary checks for the typeinfo pointer, there is now a 'Undefined' fallback type for trees, nodes and sockets. These types are used as placeholders in case the real type is not registered and can provide useful visual feedback on undefined nodes.
PyNodes opens up the node system in Blender to scripters and adds a number of UI-level improvements.
=== Dynamic node type registration ===
Node types can now be added at runtime, using the RNA registration mechanism from python. This enables addons such as render engines to create a complete user interface with nodes.
Examples of how such nodes can be defined can be found in my personal wiki docs atm [1] and as a script template in release/scripts/templates_py/custom_nodes.py [2].
=== Node group improvements ===
Each node editor now has a tree history of edited node groups, which allows opening and editing nested node groups. The node editor also supports pinning now, so that different spaces can be used to edit different node groups simultaneously. For more ramblings and rationale see (really old) blog post on code.blender.org [3].
The interface of node groups has been overhauled. Sockets of a node group are no longer displayed in columns on either side, but instead special input/output nodes are used to mirror group sockets inside a node tree. This solves the problem of long node lines in groups and allows more adaptable node layout. Internal sockets can be exposed from a group by either connecting to the extension sockets in input/output nodes (shown as empty circle) or by adding sockets from the node property bar in the "Interface" panel. Further details such as the socket name can also be changed there.
[1] http://wiki.blender.org/index.php/User:Phonybone/Python_Nodes
[2] http://projects.blender.org/scm/viewvc.php/trunk/blender/release/scripts/templates_py/custom_nodes.py?view=markup&root=bf-blender
[3] http://code.blender.org/index.php/2012/01/improving-node-group-interface-editing/
Two example files that crashed texture nodes.
- On delete texture nodes, it should free the exec cache (because this cache
stores the node pointer.
- On redo, nodes can exist can exist without typeinfo set. Exec (free) code
was not checking for that. Don't ask me why this happens... tex nodes are weird.
- Cycles materials now render in Blender Internal too, skipping the nodes.
Not very useful, but at least things then show up on renders and in
previews.
- Node editor: if wrong shader nodes are in a tree, they draw with thene
color RED ALERT headers now. (Switching render engine will show it).
for now subtype is not defined, but once we start parsing the metadata we can set texture inputs as FILEPATH
also, it takes relative strings and convert to absolute for all strings (which is arguably a good solution, but
should work for now)
Storing this list in the node has the advantage of requiring far fewer calls to the potentially expensive internal_connect callback. This was called on every node redraw ...
Also it will allow Cycles to properly use the internal links for muted nodes, which ensures consistent behavior. The previous method was not applicable in Cycles because transient list return values are not supported well in the RNA and particularly the C++ API implementation.
Problem here is that muted nodes and reroute nodes are supposed to be removed from the execution node tree during the localize function. However, this is function is apparently only used during preview renders and must be considered a hack (is there anything that is not a hack in BI?)
Now the mute/reroute check happens in the node tree exec functions still used by BI and the legacy compositor and texture nodes. It uses the same internal_connect function from nodes to assign input stack indices directly to outputs (which also avoids overhead). Localize function also still does this. Cycles/Tile should also implement muting/reroute in their intermediate node layers by using this function, then it could be removed from localize too.
It is not a well-supported feature of the primary node systems (shader, compositor, texture) in Blender. If anybody wants to create a node system that has actual use for loops, they can do so much more elegantly with Python nodes, but it does not have to be a core node type in Blender. Removing this should ease node code maintenance a bit.
- movieclip unlink didn't clear node ID pointers from the scene (leaving dangling pointers).
- mask datablock unlink was clearning references from scene nodes twice.
